Orientation and organization of the presynaptic active zone protein Bassoon: from the Golgi to the synapse

(1)

(2)

Orientation and organization of the presynaptic active zone protein Bassoon:

from the Golgi to the synapse

Dissertation

for the award of the degree

“Doctor rerum naturalium”

of the Georg-August-Universität Göttingen within the doctoral program CNMPB-GGNB

of the Georg-August University School of Science (GAUSS)

submitted by

Tina Ghelani

Born in Kolkata, 02.01.1986

March 22

^nd

2016 , Göttingen

(3)

Thesis Committee

Dr. Nina Wittenmayer

(Deptartment of Anatomy and Embrology, University Medical Center Göttingen) Prof. Dr. Blanche Schwappach.

(Department of Molecular Biology, University Medical Center Göttingen) Prof. Dr. Fred S. Wouters

(Laboratory for Molecular and Cellular Systems, Institute for Neuropathology, University Medical Center Göttingen)

Members of the Examination Board

1^st Referee: Prof. Dr. Thomas Dresbach

(Deptartment of Anatomy and Embrology, University Medical Center Göttingen) 2^nd Referee: Prof. Dr. Blanche Schwappach

(Department of Molecular Biology, University Medical Center Göttingen)

Further members of the Examination Board

Prof. Dr. Fred Wouters

(Laboratory for Molecular and Cellular Systems, Institute for Neuropathology, University Medical Center Göttingen)

Dr. Nina Wittenmayer

(Deptartment of Anatomy and Embrology, University Medical Center Göttingen) Prof. Dr. Micheal Hörner

(Department of Cellular Neurobiology, Johann-Friedrich-Blumenbach-Institute for Zoology and Anthropology, University of Göttingen)

Camin Dean, Ph.D.

(Laboratory of Trans-synaptic Signaling, European Neuroscience Institute Göttingen)

Date of oral examination: 12^th of May, 2016

(4)

Declaration

I hereby declare that this thesis entitled “Orientation and organization of the presynaptic active zone protein Bassoon: from the Golgi to the synapse” has been written independently and with no other sources and aids than quoted.

Tina Ghelani,

Göttingen, 8^th of August, 2016

(5)

Summary

Neurotransmitter release sites at the presynaptic terminus membrane, known as the active zones (AZs), are surrounded by synaptic vesicle pools and a dense network of five cytomatrix of the active zone (CAZ) proteins. At CNS synapses, Munc13s, RIMs, CAST/ELKS, Bassoon, and Piccolo predominantly form the CAZ scaffold, and together have been shown to promote short- and long-term plasticity by binding to Ca²⁺ channels and enabling priming and docking of synaptic vesicles to the presynaptic membrane. Even though the components of the CAZ are known, how they are exactly assembled opposite postsynaptic specializations is not yet understood.

It has been shown that AZ proteins (AZPs) are transported on 80nm dense- core vesicles called Piccolo/Bassoon transport vesicles (PTVs), to synapses in aggregates together with synaptic vesicles. In addition, Golgi-derived AZ precursor vesicles that transport these proteins have been reported to take different paths out of the Golgi and carry only a small subset of AZPs, although how all AZPs reach and generate a complete and functional AZ at the presynaptic terminus is still under investigation. These observations suggest traffic of a range of different transport vesicles, carrying subsets of AZPs to synaptic sites, and indicates that the mechanisms influencing the final assembly of AZPs at the presynaptic terminus may be predetermined as early as their sorting and loading onto transport vesicles at the Golgi.

To address this hypothesis, this study, examines the localization of endogenous AZPs, using super-resolution microscopy, for the first time at Golgi substructures, the soma, and in the developing axons of hippocampal neurons. AZPs are specifically localized at and around their respective Golgi lamella, in a range of signal sizes that correspond to different loaded transport-carrier types, and present low co-localizations, with one another, in developing axons. This distribution signifies the importance of early sorting and loading of preassembled AZP subsets in the soma. In order to understand the underlying mechanisms that dictate the specific localization of CAZ proteins, a detailed study of the nanostructural orientation and organization of AZPs, at different cellular locations, is required, but hampered by the limitations imposed by the use of primary and secondary antibodies. To overcome this technical limitation, I introduce, characterize, and use new full- length second-generation Bassoon constructs that are optimized, with respect to their targeting behavior in neuronal cells, and are endowed with tags that can be detected with very small camelid antibodies, so called nanobodies, for super-resolution imaging.

Bassoon is one of the largest CAZ proteins and among the first AZPs to be incorporated at young synaptic sites. It is known to bind to other AZ proteins

(6)

in the CAZ scaffold, and provides structural stability to the CAZ scaffold by downregulating local ubiquitination. This suggests a central role of Bassoon in CAZ scaffold generation. In addition, Bassoon is also the mammalian AZP with the largest cohort of mutant and full-length constructs available, making it an ideal candidate for this study.

STED and FLIM imaging show that full-length Bassoon molecules possess an open and extended conformation at the TGN and are organized 6—20nm from the TGN with neighboring N-termini of molecules in close proximity to each other. Further, these studies show that the first 94 amino acids of Bassoon’s N-terminus, but not its myristoylation motif, determines its correct subcellular localization to the TGN, while Bassoon’s CC2 domain is sufficient for recruiting the protein to the Golgi. A novel conformation change is observed as the Bassoon molecule travels from the Golgi to synaptic sites, where the molecule appears to lose its extended conformation during trafficking on ChromograninA-positive PTVs, and returns to its extended orientation at synaptic sites. Within these sites, in CAZ scaffold, Bassoon molecules have been previously shown to be oriented with their N-termini extending 80nm into the presynaptic bouton and their C-termini positioned around 50nm from the presynaptic plasma membrane. In this study I show that the N-termini of neighboring Bassoon molecules are organized in close proximities of ≥5nm from each other. This result suggests that the organization of Bassoon molecules within the CAZ scaffold closely resembles triangular dense projections regularly observed in EM images of CNS presynaptic sites. Therefore the orientation and the organization of Bassoon molecules promotes structural stability by inhibiting localized ubiquitination and forms the backbone that other AZPs bind to within the CAZ scaffold.

In conclusion, the data reported here suggest that the orientation and organization of Bassoon molecules plays an important role in promoting local subcellular mechanisms from influencing its localization and sorting at the TGN to providing structural stability to the CAZ scaffold, while its change in conformation on its journey, to the CAZ, highlights the first step in understanding the sequence of mechanisms involved in mammalian AZ assembly and synapse maturation.

Keywords: Bassoon; orientation; super-resolution imaging; AZ assembly;

cytomatrix of the active zone; STED; FLIM

(7)

List of Figures

Figure 1: Active zone ultrastructures of vertebrate and invertebrate neurons. .... 18 Figure 2: Ultrastructural and schematic organization of AZPs and SV pools of

central synapses.. ... 25 Figure 3: Bassoon, Piccolo and Munc13-1 associate with specific Golgi sub- compartments. ... 52 Figure 4: Distribution and localization of AZPs to their respective Golgi sub- compartments with and without a 19°C block using STED microscopy. . ... 54 Figure 5: Size populations of endogenous Bassoon, Piccolo, and Munc13-1

signals at their Golgi sub-compartments. ... 58 Figure 6: Early localization of recombinant Bassoon and Munc13-1 to dense-core

vesicle marker: Synaptotagmin4 (Syt4). ... 59 Figure 7: Distribution of Bassoon, Piccolo and Munc13-1 in developing neurons

with epifluorescence microscopy. ... 61 Figure 8: Distribution of Bassoon, Piccolo and Munc13-1 during high neuronal

trafficking using STED microscopy ... 62 Figure 9. A schematic of the second-generation Bassoon constructs used in this

study. ... 65 Figure 10: Recombinant Bassoon localizes closely with the TGN38 in young

neurons. ... 67 Figure 11: Double-tagged recombinant Bassoon colocalizes with synaptic

markers in adult neurons. ... 69 Figure 12: Single-tagged recombinant Bassoon also colocalizes with synaptic

markers in adult neurons.. ... 70 Figure 13: Mutation of the myristoyl group of Bassoon does not impede the

normal Golgi and synaptic localization of the protein in young and adult neurons. ... 72 Figure 14: Using GFP and RFP specific nanobodies to visualize of single

molecules of tagged Bassoon protein under 30nm resolution. ... 74 Figure 15: Super resolution localization of GFP tagged Bassoon (95-3938) and

full-length Munc13-1 at their respective Golgi sub-compartments. ... 76

(11)

Figure 16: Full-length Bassoon localizes specifically to the trans-Golgi network compartment instead of the trans-Golgi sub-compartment. ... 78 Figure 17: Orientation of mRFP-Bsn-mEGFP at the trans-Golgi network. ... 82 Figure 18: Orientation of mRFP tagged full-length Bassoon constructs at the

trans-Golgi network (TGN). ... 84 Figure 19: Orientation of single and double-tagged full-length Bassoon constructs

at the trans-Golgi network marker: Syntaxin 6 (Syn6). ... 86 Figure 20: FLIM imaging of N-terminus of Bassoon to TGN38 to visualize

interaction within 5nm. ... 88 Figure 21: A summary of the orientation of full-length Bassoon molecules at the

trans-Golgi network. ... 90 Figure 22: Orientation of first-generation mutant Bassoon constructs at the trans- Golgi network. ... 92 Figure 23: Comparing the orientation of the double-tagged myristoyl mutant

Bassoon construct to the double-tagged full-length the trans-Golgi network.

... 95 Figure 24: Orientation of G2A-mRFP-Bsn-mEGFP myristoyl mutant construct in

endogenous Bassoon-free Bsn^–/– knockout mice and their Bsn^+/+wildtype littermates. ... 97 Figure 25: Orientation of full-length single-tagged Bassoon constructs to

chromogranin A (CGA) positive PTVs in the soma and down the axon. ... 99 Figure 26: Organization of neighboring Bassoon molecules at the soma. ... 103 Figure 27: Organization of Bassoon molecules at synaptic sites. ... 105 Figure 28: A diagrammatic representation of AZPs localizations to different Golgi

subcompartments. ... 110 Figure 29: Structure and orientation of Bassoon and Piccolo proteins at the

presynaptic terminus. ... 117 Figure 30: A diagrammatic representation of the orientation and organization of

tagged Bassoon molecules.. ... 125

(12)

List of Tables

Table 1. Interaction partners of Bassoon. ... 20

Table 2. Primary antibodies List ... 32

Table 3. Seconday antibody list. ... 34

Table 4. DNA plasmids list ... 36

Table 5. Sequencing primers. ... 42

(13)

Abbreviations

AZ: Active Zone

AZP: Active Zone Proteins

CAZ: Cytomatirx of the Active Zone PSF : Point Spread Function FWHM : Full Width at Half Maximum

STED : Stimulated Emisson Detection Microscopy STORM: Stochastic Optical Reconstruction Microscopy FLIM : Fluorescence-lifetime imaging microscopy FRET : Förster resonance energy transfer CGA: ChromograninA

Syt4: Synaptotamin4 a.a. : Amino acid CC2: coiled-coil 2

BsnGBR : Bassoon Golgi Binding Region KO : Knock-out

WT Wildtype LUT: Look up table

GFP : Green fluorescence proteins RFP : Red fluorescence proteins SV: Synaptic vesicle

DCV: dense-core vesicles PM: Plasma membrane DP: Dense projection EM : electron microscope SVs : synaptic vesicles DCVs: dense-core vesicles

PTVs: Piccolo-Bassoon transport vesicles neuromuscular junctions (NMJ)

RIM: Rab-interacting molecules VGCCs: voltage gated calcium channel

(14)

(15)

Introduction

Chapter 1 Introduction

Synapses are specialized sites of contact between two neurons, designed for rapid communication via the chemical release of neurotransmitters from the presynaptic neuronal partner to the postsynaptic neuron. The term synapse was first coined in 1897 by Sir Michael Foster and Sir Charles Scott Sherrington, and based in part on both popular and contradictory theories of nervous system organization¹, of the time, i.e. the Reticular model (supported by Camillo Golgi’s work) and the Neuron Doctrine (based on Santiago Ramón y Cajal’s findings).

Both forefathers of modern neuroscience performed histological studies of subsets of neurons and their observations were formulated into the two models². The reticular model postulated by Joseph von Gerlach in 1872, describes the nervous system as a continuous syncytial reticulum consisting of nerve fibers, dendrites, and neurons, nourished through their cell somas, and directly connected to each other over cytoplasmic bridges³. While the Neuron Doctrine, formulated by Waldeyer-Hartz in 1891, states that the nervous system is not a continuous reticulum of tissue, but rather consists of separated discontinuous units or cells that became known as neurons. These cells were described as consisting of three main subcellular areas: the soma, fine tree-like processes known as dendrites, and a single long axon⁴. This model was further added to by the law of dynamic polarization, that states that neuronal signals only travel in one direction in neurons i.e., from dendrites, through cell bodies, down axons to the synapse⁵.

(16)

Introduction This was an exciting time in neuroscience history where based on purely structural histological data, the basic understanding of the nervous system was deciphered. The next decade saw a range of physiological and biochemical experiments that described the chemical nature of synaptic signal transmission, but it was not until after the development of the electron microscope (EM) in 1933, that the fine structure of synaptic organization was revealed. The first EM images of the synapses were taken in 1953⁶, followed very shortly thereafter by a range of EM studies showing electron-dense regions at the pre- and postsynapse, and a synaptic cleft separating these structures. This confirmed the Neuron Doctrine, and the presence of secretory vesicles, now known as synaptic vesicles (SVs), in the presynaptic terminus (Figures 1D and 2A). These vesicles were postulated to contain neurotransmitters and explained the previously described quantal release of neurotransmitters at the synapse observed by Sir Bernard Katz and colleagues at about the same time in the early 1950s⁷.

Although the growing number of EM studies provided a wealth of new information about the ultrastructure of the synapse, these alone were unable to identify the molecular components and proteins that constituted the electron-dense regions on both sides of the synapse.

The last 50 years have seen the application of genetic, biochemical, molecular biological and genomic methods to uncover a large number of proteins, in different model organisms, that constitute the composition of the synaptic apparatus. The synaptic proteins discovered included a range of secretory vesicles transporting synaptic vesicle proteins, peptides involved in vesicle docking and fusion, receptors of neurotransmitters, ion channels, enzymes that regulate processing, and of course the neurotransmitters themselves to the plasma membrane. In addition a range of extracellular matrix proteins, cellular signaling proteins, cell adhesion molecules that promote neuronal contacts, cytoskeletal proteins that form the backbone structure of synaptic apparatus and a number of scaffolding proteins that help mediate the structural organization of the different classes of proteins on both sides of the synapse were identified². This combination of proteins equips the presynaptic terminal for regulated depolarization and calcium-dependent exocytosis of neurotransmitter from synaptic vesicles, and the postsynaptic terminus for neurotransmitter detection by clustering of neurotransmitter receptors. Synapses can be found along the axon of a neuron at presynaptic sites known as en passant boutons or at the axonal distal end known as boutons terminaux⁸.

(17)

Introduction The active zone (AZ) and its CAZ scaffold These boutons contain resting and recycling pools of synaptic vesicles, as well as a network of cytoplasmic scaffolding proteins known as cytomatrix of the active zone (CAZ) that tether and organize proteins and vesicles required for neurotransmitter release, around roughly 500nm-long specialized regions on the plasma membrane known as the active zone (AZ). The presynaptic membrane of AZs possesses synaptic vesicle sorting and fusion machinery proteins, and is the site of neurotransmitter release.

1.1. The active zone (AZ) and its CAZ scaffold

The CAZ scaffold of presynaptic proteins is observed as the electron-dense structures in EM images, and this structure is organized exactly opposite the postsynaptic scaffold. The most important function of the CAZ is to restrict neurotransmitter release to AZs and regulate its properties. The CAZ scaffold organizes and separates neighboring AZs and vesicle docking sites while providing a protein network for proteins lacking transmembrane regions, which cannot integrate into the presynaptic plasma membrane, to localize in close proximity to their binding partners and the plasma membrane^9,10. The CAZ functions to support the structure and function of the AZ site by modulating synaptic vesicle pools by influencing their recruitment, priming, and docking¹¹. In addition, its components have also been shown to regulate the organization of Ca²⁺-channels in the plasma membrane, which indirectly influences and connects the Ca²⁺ influx within the terminus and to the balance of exo- and endocytic events of the readily-releasable pool (RRP) of synaptic vesicles¹².

Different AZ sizes, organizations and CAZ protein compositions exist and have been studied in detail in invertebrate Caenorhabditis elegans (C. elegans), Drosphila melanogaster, and vertebrate mice and rat animal models. CAZ composition is species-specific, although AZ size and the CAZ organization in these AZ is influenced by the synapse size, its morphology and the propensity of its function. For example, neuromuscular junctions (NMJ) are large synapses that possess thousands of SVs at elaborately organized AZs and PSDs and are programmed to ensure precise and reliable signal inputs for muscle contractions to occur with high fidelity. On the other hand, central nervous system synapses, such as glutamatergic hippocampal neurons have smaller AZs and less robust neurotransmitter signaling, which allows these neurons to modulate their functional propensity and allows these neuronal networks to have synaptic plasticity^10,13,14.

The central synapses have simple AZ structures characterized by electron–dense projections (DPs) that define their CAZs, which connect docked SVs and the readily releasable pool (RRP) of SVs via fine filamentous projections and tether them close to release sites on the plasma membrane. Vertebrate NMJ AZs share are similar to drosophila NMJs and are large structures with a linear organization of SV layers that are orchestrated through different filamentous structures, which

(18)

Introduction The active zone (AZ) and its CAZ scaffold tether the vesicles to the PM. Drosophila NMJ AZs are also called T-bars, describing the meshwork of filament tethers that stretch out to form a consolidated dense site on the plasma membrane into a horizontal platform that organizes the SV layers. C. elegans NMJ AZs although large in size have an AZ structure similar to classical vertebrate central synapses. The most complex vertebrate AZ structures are seen at photoreceptor cell ribbon synapses. These synapses have an elongated AZ structure at the PM, known as an “archiform density”, which is attached to a long horizontal filament known as the synaptic ribbon. The ribbon tethers SVs all around it and brings them in close vicinity to the AZ site¹³ (Figure 1).

Figure 1: Active zone ultrastructures of vertebrate and invertebrate neurons. Taken from Ackermann et al, 2015 shows the different AZ structures of NMJ, photoreceptor, and central nervous system synapses of vertebrates and invertebrates. The AZ structures show the arrangement of dense-electron projections of the CAZ proteins and the SVs at presynaptic termini of these different synapses.

Vertebrate AZs and invertebrate AZs have many features and variants of core CAZ components in common. Although a greater molecular diversity of vertebrate CAZ proteins, in the form of splice variants and gene duplications, is present to support the functional diversity of different types of AZs and the modulation of synaptic plasticity in a network of central nervous system (CNS) neurons.

Five prominent core CAZ proteins have been identified in the vertebrate synapses namely Bassoon and Piccolo, Rab-interacting molecules (RIMs), Munc- 13s and ELKSs/CAST¹¹. The invertebrate Drosophila genome has similar orthologs to vertebrate CAZ proteins. For instance drosophila CAZs have Fife (a Piccolo homolog), Bruchpilot (an ortholog of CAST), DUNC-13 (replaces Munc- 13), and DRIM (Drosophila RIM)^15–18. A Bassoon homolog has not yet been identified in Drosophila. Similarly the CAZ composition of C. elegans, comprises of SYD-2, Liprin-a, ELKS-1, UNC-10/RIM, and UNC-13 (Munc-13 variant)^19–21.

(19)

Introduction Role of CAZ scaffold protein Bassoon

1.1.1. Role of CAZ scaffold protein Bassoon

Bassoon is the second largest CAZ protein, weighs ~420kDa and forms a, multi- domain protein. It was first identified in a rat cDNA screen of synaptic junctions and found to localize in the synaptosomal and synaptic junctional fractions but not in the soluble and myelin-sheath protein fractions. The Bassoon gene is localized on chromosome 9F, possess a 13kb coding sequencing which consists of 10 exons, with more than half of it sequence encoded from exon5 ²². Bassoon is one of three vertebrate specific presynaptic proteins, and coincidently shares a large amount of homology, in the form of homology domains, with one of the other two remaining vertebrate specific AZ proteins (AZPs), that is also a key CAZ protein; Piccolo^22,23. In silico predictions have estimated Bassoon and Piccolo possess an 80nm stretched-open, filamentous structure that is rich in prolines and glycines, which promotes their structure, although both proteins have several highly compact regions in their structure²⁴. These compact regions form the two N-terminal zinc-finger domains, three coil-coil domains that Bassoon and Piccolo share in homology but use to interact with various secretory, transport, and synaptic proteins. In general, the N-terminus of Bassoon contains the zinc fingers that inhibit the local ubiquitination activity of seven in absentia homolog 1 (Siah1) and promotes synaptic stability²⁵. The central CC2 domain region promotes sorting and transport regulatory mechanism as it is flanked by a CTBP binding site (may be involved in Bassoon sorting at the TGN and balancing its expression)^26–28, and a dyein-light chain binding site (mediates retrograde transport of Bassoon and Piccolo vesicles)²⁹, while the CC2 domain itself is the oligomerization site of Bassoon and Piccolo molecules and might promote assembly of AZPs at either the Golgi or at the AZ^30–32. On the other hand, the C- terminal region of Bassoon possess binding sites for a large range of synaptic AZPs such as CAST^26,31,33, Munc-13³¹, RIMs³⁴, RIM-binding proteins³⁴, voltage gated calcium channels (VGCCs)³⁵, enzymatic activity regulator D-amino acid oxidase³⁶, and SV protein Mover^37–39. The interaction partners of Bassoon are in more detail described in Table 1.

Through these multi-domain binding sites and its extended structure Bassoon integrates into various mechanisms at play in the presynaptic terminus. These mechanisms include, organization of neurotransmitter release sites by influencing calcium channel localizations, bringing CAZ scaffold proteins and SV pools in close proximity to these sites, maintaining the structural stability, and synaptic integrity, while modulation the local synaptic and synapto-nuclear signaling pathways at play around the AZ⁴⁰.

To understand the exact cellular mechanisms that Bassoon is involved in that regulate presynaptic transmission, a range of studies have been performed using deletion mutants of Bassoon. A partial deletion mutant Bsn^Δ^Ex4/5 generated by deleting most of exons four and five of Bassoon, did not influence synaptic

(20)

Introduction Role of CAZ scaffold protein Bassoon transmission of intact cultured hippocampal neurons, although a larger number of silent synapses were generated in Bsn^Δ^Ex4/5 mice²². Similar results were observed in another study that performed a shRNA-mediated knockdown of Bassoon in autaptic hippocampal cultures^23,41. Studies of the Bsn^Δ^Ex4/5 mutant in the high throughput synapses of cerebellar mossy fibers showed that Bassoon is involved in synaptic vesicle replenishment. This mutant and a full knockout of Bassoon showed that the loss of Bassoon caused slowed vesicle recycling and produced a stronger depression in during high-frequency signal transmission⁴². In addition, Bassoon deficient photoreceptor and inner hair cells neurons present a dramatic loss of ribbon synapses and their associated proteins, similar to the loss of SV around T-bars of Bruchpilot deficient drosophila NMJ^43,44. In inner hair cells, the loss of ribbon synapses, that float into the cytosol, of Bassoon deficient mice severely affect the synchronous compound activity of the auditory nerve;

therefore Bassoon is essential for normal hearing in these mice⁴⁴. These results show that Bassoon molecules may behave as tethers for SVs in different AZs and/or that they emphasize the role of vesicle tethers in regulating vesicle replenishment⁴⁵.

Since its discovery in 1998, the extensive number of Bassoon studies show that there is no clear unifying role of Bassoon in vertebrate central, NMJ and sensory ribbon synapses, as Bassoon seems to have different roles in these synapses, although it is clear that irrespective of its exact role in the different AZs it promotes the presynaptic activity at specialized AZ sites⁴⁶.

(21)

Introduction Role of CAZ scaffold protein Bassoon

Table 1: Interaction partners of Bassoon. Modified from Gundelfinger, 2016. Abbreviations: Bsn, Bassoon; Pclo, Piccolo; CC1–3, coiled-coil regions; mouse: (ms) Znf: Zinc-finger domain

Bassoon binding site

Interaction partner

Description/

Potential function Reference

Cellular process Pclo, Znf1

521–582 Pclo, Znf2 1010–1071

Bsn, Znf1 162–225 Bsn, Znf2 459–523 (rat)

Siah1, seven in absentia homolog 1

E3 ubiquitin ligase, ubiquitinates SV proteins, component of the ubiquitin- proteasome system.

Waites et al.

(2013)

25

Protein turnover/

degradation

Bsn aa 1360–1692

(rat) , not present

in Pclo

Dynein light chains Dlc-1, Dlc-2

Link to dynein motors, (retrograde) transport of Piccolo- Bassoon transport vesicles

Fejtova et al. (2009)

29

Membrane trafficking

Bsn CC2 domain

2088–2563 (rat)

Bsn CC2 Pclo CC2 (aa 3094–3218

mouse)

Homo-/hetero- dimerization region, presumably scaffold formation,

Golgi-binding domain of Bassoon

Dresbach et al. (2006) Wang et al.

(2009) Maas et al.

(2012) ^30–32

Scaffolding and Assembly of CAZ core complex Bsn, CC3(rat)

2933–2995 Bsn, CC3(ms) 2873–3077 Pclo, CC3 (rat) 3601–3960 Pclo, CC3(ms)) 3657–3715

ERC2/

ELKS2/

CAST

Interaction with CAZ scaffolding proteins.

Potentially involved in anchoring synaptic ribbons to the active zone

Takao-Rikit- su et al.

(2004), tom Dieck et al.

(2005), Wang et al.

(2009)^26,31,33

Scaffolding and Assembly of CAZ core complex

Bsn, Ser2845

(rat)

14–3–3𝜂 (and other

isoforms)

Phospho-dependent regulation of anchoring of bassoon to CAZ Phosphorylation depends on

RSK family

Schröder et al. (2013) Cellular signaling, Scaffolding and Assembly of CAZ core

complex? Phospho-dependent

regulation of anchoring of Bassoon to CAZ,Phosphorylation depends on RSK family

Schröder et al. (2013)

47

Cellular signaling, scaffolding and potentially assembly of CAZ core complex Bsn, aa

2715–3263 (rat) Not tested

for direct interaction

with Pclo

D-Amino Acid Oxidase,

Enzyme metabolizes the NMDA receptor co-agonist D-serine.

DAO activity

significantly inhibited by interaction with Bsn

Popiolek et al. (2011)

36

Cellular signaling, Regulation of enzymatic activity

Bsn, aa 3601–

3820 (C-term) (mouse)

Munc13 (N-term) and RIM

Interaction with presynaptic scaffolding

Wang et al.

(2009)

31

Scaffolding and Assembly of CAZ core complex, SV priming Bsn aa 3263–

3938 (rat) Absent in Pclo

Mover/TPRGL SV protein, negative regulator of synaptic release probability

Kremer et al.(2007), Ahmed et al.(2013), Korber et al.

(2015)^37–39

Potential role in membrane trafficking Regulation of exocytosis

(22)

Introduction CAZ scaffold protein Piccolo

1.2. Members of the CAZ scaffold in central synapses 1.2.1. CAZ scaffold protein Piccolo

Piccolo (also known as Aczonin) is the largest CAZ protein in the scaffold, with a molecular weight of ~550 kDa, and it shares with Bassoon a high degree of sequence homology (50 – 80 % common sequence identity), at two N-terminal zinc-fingers and three coil-coil domains together known as the Piccolo-Bassoon homology domains. These domains enable Bassoon-Piccolo interactions at the central CC2 and allow Bassoon and Piccolo complexes to compete for binding with other synaptic and scaffold proteins²⁶. For example the N-terminal zinc finger domains of Piccolo and Bassoon bind to Siah1²⁵, though it is not yet understood whether the proteins collaborate or compete to inhibit Siah1 activity, while the C-terminal CC3 domain of Bassoon and Piccolo have been shown to competitively interact with the CC2 domain of CAST⁴⁸. Piccolo in addition has some unique interaction domains that separates its function from Bassoon’s function and contributes to roles that CAZ scaffold provide towards enabling presynaptic transmission.

The N-terminal zinc-finger domains of Piccolo additionally have 40 % and 39 % homology to the zinc- finger domains of rabphilin-3A and RIM respectively, and contain a PRA1 binding site that has been shown to mediate Piccolo interactions with rab3A and VAMP2 receptors, thus linking Piccolo and synaptic vesicles pools at the AZ⁴⁹. Piccolo possesses a few proline-rich sequences over its structure. A proline patch in the N-terminus of piccolo interacts with actin binding protein 1(Abp1), which links Piccolo by interacting with actin, dynamin⁵⁰, and a GTPase that mediates the fission of SV vesicles⁵¹.

The central region of Piccolo also possess another proline patch that interacts with profilin, an actin-binding protein, that influences the actin-dynamics within presynaptic terminal⁵². This enables Piccolo to link the CAZ scaffold of proteins to cytoskeleton of the presynaptic terminus as well as incorporate itself into the process of endocytic fission at the AZ site.

Piccolo possesses two C2 domains, namely C2A and C2B domains, within its C- terminus that are not present in Bassoon. The C2A domain has been shown to undergo a conformation switch, upon binding Ca²⁺, that promotes dimerization and Ca²⁺-dependent phospholipid binding implicating a role for Piccolo in short- term plasticity⁵³. Overall, by virtue of its large size and its CAZ interaction partners, Piccolo incorporates itself into the CAZ scaffold and links the CAZ scaffold the action cytoskeleton and SV recycling events at the presynaptic terminus.

(23)

Introduction Munc-13s, RIMs and CAST/ELKs

1.2.2. Munc-13s, RIMs and CAST/ELKs

Munc13, the vertebrate specific isoform of unc-13 found in C. elegans, is a small yet integral CAZ protein that interacts and neuronal Rab3 isoforms (RIMs) to regulate presynaptic neurotransmitter transmission^54,55,56. CAST/ELKs are quintessential CAZ as a scaffold proteins that despite having different roles in different synapses are always involved in organizing the CAZ structure¹⁸. All three proteins bind to specific C-terminal subdomains of Bassoon and Piccolo and are therefore integral members of the CAZ scaffold31,34,26,31,48,57.A schematic depiction of these interactions can be referred to in Figure 2.

Three Munc13 genes exist in the brain Munc13-1, Munc13-2, and Munc13-3, are roughly ~222 kDa in weight and all three isoforms share three evolutionary conserved C2 domains present in the N- terminal, central, and C-terminal regions of the protein^56,58.

The C2 domains of Munc13-1, i.e. C2B and C2C, known as the MUN domain (spans from aa 859-1531), binds to the N-terminus of syntaxin and thereby integrates itself, while linking the CAZ scaffold, to the core of SNARE complex that forms presynaptic membrane machinery at AZ sites⁵⁵. This interaction suggests that Munc13-1 may mediate synaptic vesicles docking at AZs.

Additionally, the MUN domain, when expressed in hippocampal neurons lacking Munc13s, rescues the SV priming deficits⁵⁹. SV priming is a maturation step that occurs between the vesicle docking and SNARE-mediated SV fusion steps.

The presence of SV priming is attributed to the combination of opposing electrophysiological and ultrastructural observations noted in Munc13-1 knockout and Munc13-1 and Munc13-2 double knockout mice cultures. Glutamatergic hippocampal neurons of Munc13s, upon inspection for ultrastructural deficits, showed that docked SV vesicles numbers remained unchanged, despite the complete loss of evoked EPSCs in synapses deficient for Munc13s^60,61. These results suggest Munc13s use their conserved MUN domain to play a crucial role for synaptic vesicle maturation.

The role of Munc13s in SV priming is compounded by the interaction of Munc13- 1’s C2A domain with N-terminal zinc-finger domain of RIM. Disruption of this interaction in RIM deficient synapses prevents synaptic vesicles from reaching and fusing with the presynaptic membrane, as the lack of RIM forces Munc13-1 C2A domains to homo-dimerize thereby inhibiting Munc13’s SV priming and fusion competence^55,62. This observation suggests that the N-terminus of RIM is essential for activating Munc13-1 molecules to mediate SV priming.

The RIM protein family consists of neuronal Rab3 isoforms that regulate neurotransmitter release at the presynaptic terminus. RIMs were first isolated and identified in a yeast two-hybrid screen of Rab3C against a rat brain cDNA library,

(24)

Introduction Munc-13s, RIMs and CAST/ELKs and found isolated on the synaptic vesicle fraction. The protein is 1554 amino acids long and consists of an N-terminal zinc finger domain comprised of alpha helices, alanine and proline-rich patches, a postsynaptic density domain, and C- terminal C2A and C2B domains. The RIM protein sequence is well conserved in invertebrates, but in vertebrates at least four RIM genes exist, RIM1—RIM4, with two isoforms each for RIM1 and RIM2⁵⁴.

The N terminal alpha helical domain of RIM binds to synaptic vesicle component rab3, forming a tight link between RIMs and SVs⁵⁵. In addition the central PDZ domain of RIM proteins binds to N- and P-/Q-type Ca²⁺channels. The interaction RIMs and Ca²⁺channels is necessary for recruiting and organizing Ca²⁺channels at the AZ. Deletion of the PDZ domain of RIMs not only affects the clustering of Ca²⁺channel at the presynaptic membrane, but causes a decrease in Ca²⁺influx, which in turn reduces the priming of SVs and thereby fails to interact Ca²⁺

channels to synaptic vesicles⁶³.

CAST or CAZ-associated structural protein, and ELKS (named after its highest a.a. content glutamate (E), leucine (L), lysine (K), and serine (S)) are CAZ proteins that are 957 a.a. and 948 a.a.-long, respectively, and share a 71 % a.a.

identity⁶⁴. Two isoforms for ELKS: ELKS𝜀 and ELKS𝛼 exist. ELKS𝛼 is the neuron specific isoform that is involved in CAZ organization, while ELKS𝜀 are ubiquitously expressed, do not localize to the CAZ, and have been implicated in GTP-dependent Rab6 interaction that mediates its secretory traffic to membranes^64,65.

CAST and ELKS𝛼 colocalizes with Bassoon in neuronal cultures and possess four coil-coil domains and a C-terminal IWA conserved domains^33,64.The first two coil-coil domains (spanning 680 a.a. of their N-terminus), are essential for CAST and ELKS𝛼 targeting to the CAZ, while the CC2 domain of CAST competitively binds to the CC3 domains of CAZ scaffold proteins; Bassoon and Piccolo⁴⁸. Although knocking out CAST does not yield an effect on synaptic transmission in their central excitatory synapses, it however crucially impairs retinal ribbon synapse transmission. Knockout of the CAST protein produces smaller AZs and diminished transmission in excitatory neurons⁶⁶, and enlarged resting SV pools in inhibitory neuronal terminals, suggesting that CAST influences SV priming⁶⁷. Together these five AZPs build a network of proteins localized in close proximity to the AZ site, link the site to SVs and prime them, arrange Ca²⁺ channel to the AZ site, promote the structural stability of scaffold by inhibiting local degradation mechanisms and link the scaffold to the dynamically changing actin cytoskeleton in the presynaptic terminus (Figure 2).

(25)

Introduction CAZ Assembly

Figure 2: Ultrastructural and schematic organization of AZPs and SV pools of central synapses.

Modified from Gundlefinger et al, 2016. (A) Cryo-electron micrograph rat central excitatory synapse (originally published in Rostaing et al., 2006). (B) Schematic organization of SVs pools within presynaptic boutons, the reserve SV pool tethered via synapsin and the docked SV pool are implanted in the CAZ (red tethers). (C) Schematic of CAZ proteins, interaction partners guiding the SV clustering, translocation, docking, priming and fusion at the presynaptic terminus.

Scale bar A, 200nm.

1.3. CAZ Assembly

Studies following the discovery of AZPs of CNS synapses initially proposed that a complete complement of AZPs present on a single, 80nm in diameter, dense- core vesicle (DCV), defined at the Piccolo-Bassoon transport vesicle (PTV), delivers the entire CAZ scaffold to the presynaptic terminus possibly by fusing with the plasma membrane. And evidently so, as an exhaustive set of synaptic proteins (such as VGCCs, SNAREs: syntaxin-1 and SNAP-25, and N-cadherin) and active zone core components Bassoon, Piccolo, RIM1, Munc13, CAST1, and CAZ-associated protein Munc18-1 were isolated in the same insoluble synaptosomal fraction of rat light-brain fractions^68,69.

Today this view of the PTV model is considered an oversimplification, as recent studies illustrate that the model is a lot more complicated. The most recent study addressing this topic shows that, in fact, the assembly of AZPs may already begin at the trans-Golgi Network (TGN) sorting compartment of the Golgi. This study also highlighted differential sorting of AZPs, as early as their localization to the Golgi, and shows that a subset of Bassoon, Piccolo, and ELKs leave the Golgi on precursors generated at the TGN, while Munc13 is localized to cis-Golgi, and RIM1 is diffused in the cytosol, only to be recruited later during a post-Golgi step³². This suggests that preassembly of AZPs occurs in a multistep process, the order of which is as yet unclear, although it may involve a maturation step in

(26)

Introduction CAZ Assembly which PTVs recruit other AZPs on their way to the AZ. The preassembly of AZP subsets, for example for Piccolo-Bassoon-ELK precursors, have been suggested to first assemble at the TGN, and require not only localization and binding to specific Golgi substructures, but an association to each other, and membrane fission and translocation regulators such as CtBP1/BARS30,32,69,70

In addition, ultrastructural studies of young hippocampal axons have revealed Bassoon and Piccolo in transit on ~220nm x 130nm transport aggregates, consisting of 1—2 PTVs and 5—6 SVs that co-traffic PTV proteins and SV proteins (such as VAMP2, synaptotagmin, synapsin-1, and SV2), together as a preassembled complex⁷¹. This clarifies the presence of synaptic proteins on synaptic vesicles and AZPs like Bassoon and Piccolo on PTVs that get isolated in the same biochemical brain fraction, whilst being loaded on different carriers. It also suggests that a certain degree of preassembly may be promoted during transport on such transport aggregates, and that the delivery of such an aggregate may suffice generation of a functional AZ site.

The transport of AZPs has been shown to require microtubule-based transport involving specific retrograde and anterograde transport motors. So far three AZP- transport complexes have been identified, namely Piccolo-Bassoon-ELKs on TGN precursors, Munc13 on cis-Golgi precursors, and RIM-Neurexin-CASK- VGCCs on unclassified precursors, that gets associated in a post-Golgi step^3,32,71. In combination with their distinct transport subgroups, AZPs may also be differentially trafficked based on the motor proteins and adaptor binding sites they possess. For example, Bassoon has been shown to directly interact with dynein- light chain (Dlc1/2) that mediates its retrograde transport, while PTVs have also been reported trafficking in an anterograde fashion. Anterograde transport of PTVs is mediated via the direct interaction of the PTVs to syntabulin, a kinesin adaptor protein^29,72. Munc13 and RIM invertebrate-specific isoforms on the other hand have been implicated in an unidirectional kinesin mediated transport²¹. Such variability in transport could be used by the neuron to implement a temporal and sequential traffic of all AZPs to the AZ. Additionally, the incorporation of two to three PTVs, each carrying a unitary load of AZP, has been shown to suffice the generation of a new AZ, which might also promote sequential acquisition of the different AZPs into the CAZ scaffold⁶⁹.

How mammalian AZs are assembled is as yet unclear, although the classical PTV model hypothesizes delivery of a complete CAZ scaffold on top of a PTV that fuses with the plasma membrane, and leaves the CAZ proteins behind, at a defined AZ site, to form a quick CAZ scaffold.

(27)

Introduction High-resolution imaging of AZPs This notion is supported by the presence of visible SV recycling events at the new AZ sites, within a matter of 30 minutes from AZ formation, suggesting that such a quick assembly of AZ sites is fashioned from the delivery of preassembled AZP sets⁷³. In addition, it might be possible that a large and single step delivery of preassembled PTV and SV proteins might initiate the formation of a functional mammalian AZ, which is unique to only vertebrate synapses, as it involves the delivery of two of the three vertebrate-specific AZ proteins, Bassoon and Piccolo⁷⁴.

Upon delivery to nascent synapses, transsynaptic adhesion molecules Neurexins and Neuroligins signal the site of AZ on the plasma membrane, according to which the CAZ complex builds up exactly opposite the post synaptic scaffold. The mechanisms orchestrating this exact spatial organization of the CAZ scaffold are also not understood⁷⁵. Once at its correct localization in the presynaptic terminus CAZ scaffold proteins interact with each other and the actin cytosketon of the presynaptic bouton to form a dense, insoluble network, from which, an AZP once incorporated cannot be washed out, even with a Triton extraction (Table 1& ref ⁷⁰).

1.4. High-resolution imaging of CAZ proteins

Ultrastructural studies of CAZ proteins of central synapses

Since their discovery roughly two decades ago, CAZ proteins of central nervous system synapses have been extensively investigated for their function. A range of ultrastructural studies of single CAZ protein deficient synapses have reported no change in CAZ scaffold structure of central synapses41,42,60,61,63,66. Although reduction in populations of clustered SV pools around the CAZ was observed in double knockouts of Bassoon and Piccolo, RIM, and CAST deficient neurons41,42,63,66. This suggests that a single mammalian CAZ protein cannot alone cause a loss or change in CAZ scaffold structure. It is therefore essential to study the ultrastructure of all five AZPs in relation to each other, to begin understanding the mechanisms that influence assembly of the CAZ scaffold. With recent advances in new super resolution techniques, it has become easier to study the precise localization of tagged and endogenous AZPs. Two recent studies, using two different super resolution techniques, STORM and EM, have illustrated the specific localizations of N- and C-termini of Bassoon and Piccolo molecules to RIM, Munc13, and calcium channel localization in the presynaptic terminus of mature synapses. The N-termini of Piccolo and Bassoon extend roughly 80nm into the presynaptic bouton, while their C-termini, the central regions of Munc-13-1 and RIM1α, and the cytoplasmic loop of the P/Q type Ca²⁺

channels localize 20—30nm from the presynaptic plasmamembrane^76,77.

(28)

Introduction High resolution imaging of AZPs STED Microscopy With the ultrastructural ground work of matured mammalian CAZ structure available, the next questions to address are: 1) how does CAZ assembly occur, and 2) what sort of maturation events do AZPs undergo to attain their final ultrastructural localization in adult neuron. I will therefore, in this study, use STED microscopy and FLIM, and a range of full-length and mutant Bassoon constructs to understand the detailed localization, orientation, and organization of AZPs at different subcellular locations in developing neurons.

1.4.1. STED imaging

Imaging subcellular structures below the diffraction-limit of light has become a widespread technique. Super resolution microscopy using stimulated emission depletion (STED) is one such technique that alters the point spread function of traditional excitation laser beam by overlaying a second, doughnut-shaped, red- shifted laser beam (the STED beam), which suppresses fluorescence emission of fluorophores located underneath the donut, leaving only the center of the excitation lazer free. The fluorophore suppression is executed through the fundamental properties of stimulated emission, wherein a fluorophore in its excited-state interacts with a photon that possesses the energy difference between the ground and excited state of the fluorophore, through this interaction the excited fluorophore is forced to return to its ground state, before any spontaneous fluorescence emission can occur. The fluorescence of the molecules that fall in the overlap regions of the STED beam and excitation beam are switched off, leaving roughly a 20—50nm in diameter point spread function of the center of the excitation beam free to image fluorescence and provide super resolution⁷⁸.

STED microscopy applies the use of further red-shifted wavelengths, with respect to the absorption spectrum of the fluorophores imaged, during stimulated emission to avoid undesired excitation. Similarly, the two-color STED microscopy setups, used in this study, employ a single 775nm STED pulsed laser beam, on spectrally distinct red and far-red dyes that are excited and detected in an interleaved fashion to diminish their spectral crosstalk⁷⁹.

STED microscopes have been used to visualize fluorescence-labeled sub-cellular structures in unprecedented detail while permitting the use of simple sample preparation and labeling techniques that employ fluorescent tags and antibodies to visualize locations of biomolecules. These microscopes can resolve single fluorescent molecules at a 20—50nm resolution range compared to the 200nm resolution of confocal microscopes, although the use of traditional mono- and polyclonal antibodies for STED imaging have been shown to inhibit the complete labeling of molecular epitopes of a protein of interest and limit the resolution obtained from the STED microscopes, due to the large sizes of primary and secondary antibodies complexes generated⁸⁰. Alternatively the use of small, camelid antibodies, comprising of only one heavy chain, known as nanobodies

(29)

Introduction High resolution imaging of AZPs FRET-FLIM have become popular to overcome the problems encountered with application of antibodies⁸¹. However the generation of nanobodies specific for a new protein of interest is expensive and time-consuming process. The recent success of RFP and GFP specific nanobodies have reduced costs and increased the possibility of super resolution imaging of a large cohort of fluorescent tagged proteins, traditionally used for biochemical assays and live imaging^82–84.

GFP and RFP nanobodies are small, high-affinity, antibody-fragments generated from a single amino acid chain, off one of the two heavy chains, of the camelid IgG antibody molecule. These antibody fragments are folded into a ~10–15kDa epitope-binding hypervariable domain of dimensions: 1.5nm in diameter and 2.5nm in height⁸¹ and are generated to identify only one specific, three- dimensional epitope on top of a RFP or GFP molecule. These nanobodies can be bought pre-coupled to two molecules of secondary ATTO-TEC dyes each with a size of 2.5nm making the entire antibody complex roughly 5nm and allowing super resolution imaging, at 20nm resolution limit, to be accurate and uninhibited.

These nanobodies are routinely used to for tagged proteins pulldown assays⁸⁴ and single-molecule localization microscopy techniques^82,83,85.

In this study I will use full-length and mutant Bassoon constructs tagged with RFP and/or GFP to understand the detailed localization and orientation of the protein in developing neurons.

1.4.2. FRET-FLIM imaging

Förster resonance energy transfer (FRET) is a dipole-dipole interaction between a pair of fluorophores that are closely positioned within 5m of each other and possess a large spectral overlap between the emission of a donor fluorophore and the absorbance spectrum of an acceptor fluorophore. FRET between donor and acceptor fluorophores quench the fluorescence of the donor, which is proportional to the efficiency of FRET recorded. FRET imaging has been extensively used to study interactions of couplets of proteins of interest in living cells and generate molecular tools in the form of FRET sensors to record various protein activities⁸⁶.

Fluorescence lifetime imaging (FLIM) is a powerful quantitative FRET approach that measures the changes in fluorescence lifetime of the excited state lifetime of the donor fluorophore in the presence of acceptor. The average lifetime of the donor is reduced when the acceptor is in a close enough proximity that permits quenching of the donor and is therefore a direct indicative of FRET⁸⁷.

This method is built around the intrinsic property of the fluorescence lifetime of fluorophores and is therefore independent of fluorophore concentration effects, microscope optical path, and moderate levels of photobleaching, which makes

(30)

Introduction Aims the recorded reduction of the donor lifetime an extremely robust and quantitative estimate of the FRET efficiency⁸⁷.

These advantages also promote the use of tagged proteins, that are not necessarily generated as genetically encoded FRET sensors, and are ideal for the study of the same AZP constructs that are also used for STED microscopy. In addition, recent studies show that TagGFP—TagRFP pairs are superior to classical CFP/YFP FRET indicators, which have a lower spectral overlap and require narrow band-pass filters that cause a dramatic loss of emission. The TagGFP—TagRFP pair also have a 1.5-fold higher spectral overlap, compared to similarly separated band-pass filters pair of TagGFP-mCherry⁸⁷. This makes FLIM imaging of readily available GFP- and RFP- tagged AZPs, an ideal choice to studying the orientation and organization of AZPs.

1.5. Aims of this study

Little is known about the mechanisms that regulate CAZ assembly and AZ maturation of mammalian synapses. With new advances in super-resolution microscopy techniques allowing more synaptic proteins to be visualized with greater detail, and compared to classical ultrastructural studies, the information describing the detailed localizations of AZPs is slowly being stitched together.

However, a comprehensive overview of how the presynaptic scaffold is initiated, assembled, and matured at AZ sites is yet to be provided. There is recent evidence that suggests the delivery of AZPs, on PTVs to the presynaptic scaffold, is through an unclear multistep process³². It is therefore possible that assembly and organization of CAZ components may already be pre-programmed and influenced by cellular process, prior to the formation of a matured AZ site and may shape the final organization of AZPs within the CAZ scaffold. To realize this information, a detailed step-by-step localization of AZPs from their first station in the neuronal soma to nascent synaptic sites is required.

In this study, I aim to dissect the detailed localization of AZPs at different subcellular locations in developing neurons. This line of investigation should help reveal molecular mechanisms that dictate the specific localization of CAZ proteins, and improve our understanding of basic cellular process that regulate sorting, trafficking and delivery of synaptic proteins to AZs. To perform these studies, I aim to characterize new full-length Bassoon constructs, optimize them for visualization by super-resolution microscopy techniques, and evaluate them in context of known ultrastructural studies. The application of these constructs provide a timely opportunity to resolve the orientation and organization of Bassoon molecules, and highlight changes in conformation that might link CAZ structure to the underlying mechanisms of CAZ assembly.

(31)

(32)

Materials

Chapter 2 Materials and Methods

2.1. Materials

2.1.1. Antibodies

Primary antibodies

Table 2: A selection of primary antibodies used in this study.

Primary antibody

Species reactivity

Dilution

factor Company Catalog number

Bassoon guinea pig 1 to 300 Synaptic

systems 141 004 Bassoon mouse 1 to 500 ENZO

lifesystems

040 111 20 Bassoon rabbit 1 to 500 Synaptic

systems

141 003

Munc13 rabbit 1 to 300 Synaptic systems

126 103 Piccolo rabbit 1 to 200 Synaptic

systems

142 003 Mover rabbit 1 to 500 Synaptic

systems

248 003

(33)

Materials

TGN38 mouse 1 to 500 BD-

Transduction Laboratories

610899

TGN38 rabbit 1 to 500 BD-

Transduction Laboratories

610898

GM130 mouse 1 to 500 Gift from Dr.

C. Dean

Syntaxin 6 mouse 1 to 300 Abcam 12370

Giantin mouse 1 to 1000 Abcam 3726

Giantin rabbit 1 to 1000 Abcam 24586 Vamp4 rabbit 1 to 1000 Synaptic

systems

136 002

GFAP mouse 1 to 1000 Synaptic systems

173 011 Map2 chicken 1 to 6000 Synaptic

systems 188 006 Smi312 mouse 1 to 2000 Covance SMI-321R Calreticulin rabbit 1 to 1000 Abcam 2907 Ankyrin G mouse 1 to 400 Gift from

Dr. C. Dean

BDNF chicken 1 to 400 Promega G164A

Rab3a mouse 1 to 300 Synaptic systems

107111

GFP mouse 1 to 3000 Abcam 1218

GFP chicken 1 to 3000 Abcam 13970

Synaptophysin mouse 1 to 1000 Sigma Aldrich 118K 4828 Synapsin1/2 mouse 1 to 500 Synaptic

systems

106 002

Homer1 rabbit 1 to 300 Synaptic systems

160 003 SHANK2 guniea pig 1 to 1000 Synaptic

systems

162 202

PSD95 mouse 1 to 300 Abcam 99009

(34)

Materials

Secondary antibodies

Table 3: A list of Alexa Flour dyes used for light microscopy and FLIM, Atto dyes and STAR dyes used for STED microscopy and RFP and GFP nanobodies used for STED microscopy and FLIM.

Alexa Fluor®350 Alexa Fluor®488

Species

Dilution factor

Antigen binding

fragment Company

Mouse 1 to 1000 IgG Invitrogen

Rabbit 1 to 1000 IgG Invitrogen

Guinea pig 1 to 1000 F(ab‘)2 Jackson

Alexa Fluor®546

Species

Antigen binding

Mouse 1 to 1000 IgG Invitrogen

Rabbit 1 to 1000 F(ab‘)2 Invitrogen

Guinea pig 1 to 1000 IgG Invitrogen

Cy3

Species

Antigen binding

Mouse 1 to 1000 Jackson

Rabbit 1 to 1000 IgG Dianova

Guinea pig 1 to 1000 Jackson

Chicken 1 to 1000 Jackson

AlexaFluor®647

Species

Antigen binding

Mouse 1 to 1000 F(ab‘)2 Invitrogen

Orientation and organization of the presynaptic active zone protein Bassoon: from the Golgi to the synapse